On Sat, 25 Mar 2017, Paul C wrote:
I wish the world would use ipv6 enough for this to be worth doing, but
it's not going to have much benefit to you as there's almost no one
using it for smtp, from the last time I checked which was a few months
ago, google uses it perfectly, verizon too (maybe a few more cable
domains), yahoo looked like they were trying lol, website and some
services were v6 this year but smtp was not when I checked, hotmail
doesn't use it anywhere from what i can see, aol never will and almost
no self hosted mail server will have it. My guess is (unless gmail is
where most mail goes) that you might see a few percent like 1-5% of
mail ever use it. Not a bad research project or knowing v6, or if you
have other reasons, but actual sending out is just not happening any
time soon.
I don't see any big difference in IPv6 for mail servers and web pages
(different stats state IPV6 usage between 6 and 15% ATM, growing
exponentially).
Checking my current logfiles of the last few days and stripping all
duplicate entries (IP addresses or domain names indicate they are same) I
get following results (for outgoing TLS connections):
Server 1: 5 / 24 == 17% IPv6
Server 2: 12 / 137 = 8% IPv6
Note, that all of the IPv6 servers are also in the IPv4 list because of
the way postfix handles sending.
The percentage of IPv6 connections is much higher, as many mails go the
IPv6 hosts (especially because of google hosts), whereas all the others
get only little mail. Probably stats are lower for non TLS connections,
but who cares about these...
So while a suggestion not to care about IPv6 may have been valid in 2014.
It is simply wrong in 2017.
P.S. For server 2 simply counting IP addresses it is 78 / 330 = 19% IPv6.
Ciao
--
http://www.dstoecker.eu/ (PGP key available)
